I chose artificial intelligence as my next topic, as it can be considered as one of the most known technologies, and people imagine it when they talk about the future. But the right question would be: What is artificial intelligence?
Artificial intelligence is not something that just happened in 2015 and 2016. It’s been around for a hundred years as an idea, but as a science, we started seeing developments from the 1950s. So, this is quite an old tech topic already, but because of the kinds of technology that we have access to today — specifically, processing performance and storage — we’re starting to see significant leaps in AI development.
When I started the course entitled, “Foundations of the Fourth Industrial Revolution (Industry 4.0),” I got deeper into the topic of artificial intelligence. One of the differences between the third industrial revolution — defined by the microchip and digitization — and the fourth industrial revolution is the scope, velocity and breakthroughs in medicine and biology, as well as widespread use of artificial intelligence across our society. Thus, AI is not only a product of Industry 4.0 but also an impetus as to why the fourth industrial revolution is currently happening and will continue to do so. I think there are two ways to understand AI: the first way is to try giving a quick definition of what it is, but the second is to also think about what it is not.
Artificial intelligence as an imitation of humans
The definition that I’ve found useful in my own research is that AI is machines imitating humans. I believe this definition is important, because when many of us think about AI, we think of it as either being human or being better than human. I also think that’s not where we’re at today.
There are different types of intelligence. One is artificial general intelligence, or superintelligence. The key here is in the word artificial, meaning an imitation. And it tries to replicate what humans do very well, but it can do it at an even better speed and at scale.
So, if a human is given a 100-page book containing 100 photographs and is asked to circle every bicycle shown within, of course, the human will be able to do it and will do really well. Humans will probably do it with great accuracy, but they won’t be terribly quick at it. We’ll take the first picture, we’ll look at it, we’ll circle the bicycle, we’ll move onto the next picture, and so on until the 100th photograph. Humans barely have to think about it. It’s such a natural thing for us. But if you give this task to a computer today, it will do it in seconds, as it can imitate that human work remarkably faster. The speed really cannot be understated here. If you can combine the speed of processing with scale, you can start to augment what humans do, and find clever ways to do human processes, which many people have already been exposed to in the last few years. One of these is online recommendations. In Amazon’s early days, when you bought a book, you would be recommended some other books that you might like based on your preferences: People who bought this book also liked these books, so you might like them as well. This was a very clever usage of AI technology: Being able to look for patterns and make conclusions based on the conditions that humans set, and then serve them very quickly.
Suddenly, this technological application became remarkably beneficial. Another example is getting from point A to point B. Today, this is a product, again, of fast processing and storage — thus, artificial intelligence. If you’re sitting in traffic while using a GPS device on your smartphone, and there’s a better way to get home to avoid traffic, that technology is going to evaluate many different permutations. You could do it as a human, but you can’t do it as fast and you certainly wouldn’t be able to consider all the tradeoffs between various routes. Whereas computers are very good at processing at this volume and scale.
Some useful applications for artificial intelligence
So, as a domain, artificial intelligence is going to continue to show up in many aspects of our lives. The fact that we’re all getting connected now means that you don’t have to have scale processing on your device, as you are connected to a query that is initiated through some activity being shot to the cloud. Like this, you have massive amounts of processing. In fact, we think that ultimately, quantum computing could manifest through cloud services. It’s very unlikely that we’re going to have a quantum chip or quantum computer on a smartphone. I don’t say never, but I just don’t think it will happen. What we’ll have is cloud-based quantum processing for artificial intelligence.
This topic is especially cool in the healthcare domain. Think about how medicine works today. Medical practitioners go to school for many, many years, memorize a lot of information, then treat patients, get experience, and over the span of their career, become quite good at what they do. However, they are ultimately subject to the weaknesses of their own mortal existence. They can forget things; they can be absent-minded or, you know, just not connect the dots sometimes. Now, if we can equip a doctor physician with a computer to improve memory, options and optimization, the tools and the ability to provide medical aid suddenly change.
Let’s look at IBM’s AI initiative Watson combined with an oncologist treating a cancer patient, for example. Each patient is different, so the doctor wants to have as many details as possible about this type of cancer and the patient’s medical history to make the best treatment plan. An AI-augmented device produced for the doctor’s team could generate a scenario based on the data of every patient that has had this particular set of circumstances and that person’s characteristics. The patient could be a white male, 75 years old, living in a certain country and is of a certain genetic background. Suddenly, augmentation at an AI-level of scale and speed is changing the game.
So, my view is that we’re going to head into a world where there’s a lot more of that. Primary philosophical concerns for AI that will become very important in the future are, Can we ultimately displace the human so that these activities are just done by computers? Is there a role forward for humans in this AI future? The truth is, we’re finding that humans don’t really have value in a whole range of industries. So, how big and how massive would this displacement be?
Artificial intelligence for multiplanetary colonization
Think of Elon Musk’s dream to create a colony on Mars. The problem is that people going there now would have a six-month journey, they would have to take all their possessions, food, seeds, medicine and so on, and after that, they would have to start healing Mars and creating a livable environment. Maybe this could serve as a great opportunity to realize that sending artificial intelligence first would be better and more quickly prepare new colonies for human beings. I think we’ll probably send robots first. Although, there is talk right now that sending humans is helpful for getting government funding, as it turns out to be a human achievement. It’s all about that.
When we were doing the Apollo missions to send humans to the moon in the 1960s, at first, we sent robots and rockets and we landed them. But in 1969, when we landed a human, when Neil Armstrong walked on the moon, it changed everything. It was a completely different experience.
Of course, we will attempt a multiplanetary existence if we want to extend the survival of humanity. For going beyond the Alpha Centauri and other solar systems, though, we might just send AI. Even at amazing speed, traveling these distances takes a long time — thousands of years. Eventually, one of these crafts might be fortunate enough to be intercepted by other intelligent life whose first exposure to humans on Earth will be robots and computers. They won’t actually meet a person, a biological being like you and I.
So, this potential human future might actually be computers floating in the universe. It’s a more practical way for us to extend this idea. Otherwise, we would have to have multiple generations of humans born, raised and living on a spaceship.
Challenges for the development of artificial intelligence
The other possible course for AI development, which Elon Musk has also spoken about, is that maybe we have to separate human consciousness from the physical body. So your consciousness is basically uploaded to a computer and you can theoretically live forever, in a computer traveling through space. It boggles the mind. Really, it’s mind-expanding, because we don’t think in those terms at all. This is not how we think of the human experience — or existence at all — as we still have very old beliefs and religions about what it is to be human, what life is, the soul, and so on in our societies. Technology now is disrupting not only these beliefs, but also all our stereotypes, our understandings and our thoughts, which we as a species have had a long history with.
Anyway, let’s continue to see how AI is embedded in everything we do. If you have a technological device, it’s probably going to have some form of AI, which will be powered by cloud services that continue to evolve and continue to groove. So, anything we do as humans that requires access to large amounts of data, expert systems, expert device optimization, etc., are invariably going to be AI.
Going back to an earlier point, the example of an AI’s performance on a human process outpacing that of a human may allude to artificial intelligence being intelligent or smart, but that assumption would not be correct today. AIs really don’t have any “smartness,” they are simply performing artificial mimicry — a parlor trick that makes them look smart. It simply has access to a lot of data, or big data, and uses some mathematics.
There is a type of learning called trial-and-error learning, in which one “successively tries various responses in a situation, seemingly at random, until one is successful in achieving the goal.” There was once a game on Google where users could spend time identifying pictures. A picture would show up on-screen, users playing against each other would be given four options as to what the picture depicted, and then the users would select the word closest to the picture. The whole idea was basically training the computer to associate difficult photos with English language terms. What machine learning effectively does as a subset of AI is using historical improvements in its database to make better suggestions and predictions going forward.
This idea of predicting what happens is a sort of snowball: You can build new knowledge upon what you have already learned, which was also initially built upon what you had already learned. To add to this, the ability to predict gets much more precise over time, increasing our confidence and ambition with AI.
Artificial intelligence: Use cases
Today, apart from the takeoff and landing, a computer controls just about the entire flight of a commercial plane. We have a lot of confidence (or at least, the pilots do) that the plane can react to different circumstances and even know when to hand certain circumstances over to the pilots to contemplate. I say this with a little bit of ignorance, as a computer controlling takeoff and landing may already be in use. Today’s computers have enough processing power and information to make all the decisions necessary for flying a plane entirely.
So, I think about the positive sides of where we are. There’s a whole set of things that can just get better. I think a lot about healthcare, and I believe AI is something that will enhance our ability to make big leaps, big innovations and major breakthroughs in healthcare, which has plateaued for many decades and is long-due for a transformation.
We will certainly see improvements in production. We’ve got this notion of digital twins now: Making an artificial digital version or a copy of the real physical thing while putting sensors on the physical object. Let’s say again: A factory machine that produces a product and the digital version will entirely replicate what’s going on in the physical world. With the help of sensors and artificial intelligence, the machine can do things like predictive maintenance and even calibrate in real time. With this, we’ll be starting to get much more sophisticated, higher-quality tools of production.
I think we’ll see better use of energy, things like energy distribution and optimization, where AI applications could dramatically change the industry for good.
Another topic that I think is very important to humanity is our understanding of the weather and climate, which has a couple of major challenges on some remarkably complex topics that we’re still trying to figure out, particularly the climate. The computers that we have, which are based on classical computing systems, are not quite fast enough to even use the algorithms they’re provided to make predictions on what will happen. So, we’re going to need even faster processing. This is where, in my view, things get interesting — when we combine AI with quantum computing. The convergence of clever algorithms imitating humans driven by remarkable processing power can potentially start to solve some climate scenarios that are currently taking weeks and months to process on classical computers.
There’s a whole lot of other interesting areas that are emerging. One of the areas that has remarkable potential but is yet to be realized is augmented virtual or mixed realities. These provide us with a kind of heads-up display for a whole range of things in work, life and in general, as well as the spheres of gaming and global collaboration. Again, AI is going to help with better rendering and better simulation. That isn’t just a fleeting topic, by the way, it means an AI plus hardware, faster processing, cloud functionality, connectivity, even visualization tools and the screens all need improvements.
Artificial intelligence and other technologies
I think that brings me to my final point on this, which is: We shouldn’t think of AI alone. I think this is a mistake that many researchers and commentators often make when talking about the amazing benefits of AI, when we’re rather talking about software here. Combining AI with other technologies and industries is no less amazing.
So, I’d rather think about AI plus quantum, plus big data, plus blockchain, plus the supply chain, plus pharmaceuticals, plus healthcare. These are interesting prospects to think about, so convergence to me is really where the power of AI lies.
Think about the convergence of AI with blockchain, and how those technologies can be used together for the development of security, immutability, transparency and decentralization. I think of AI as an amplifier and blockchain as a powerful tool, and there is no doubt that the former will be involved in the set of technologies producing leading-edge solutions. No matter what, if blockchain is doing identity management or supply chain, or it is a very smart repository management system, AI baked into that architecture could make it better, faster and more precise. I can definitely see the intersections there.
One of the intersections of AI is with the Internet of Things. Let’s imagine a sensor that is part of the IoT: A complex, local sensor in a busy intersection that probably has a camera and perhaps another kind of environmental sensor embedded. Everything captured on the sensor in this IoT network is immediately sent to the cloud and processed, and maybe some action is sent back.
The latency issues involved, though, don’t make this massively practical today. You can either have a sensor that does something simple, or you bring the computer-processing to the sensor itself, developing what’s called edge computing. A certain level of AI takes place on the device so that only the really meaningful data is sent back over the cloud to be processed, and then some action is taken.
Not restricted to intersections, AI in the area of IoT sensors could be implemented in aircrafts, warehouses and cities, enabling very fast computation, filtering and decision-making at the edge. This creates a much faster and more efficient overall ecosystem.
Artificial intelligence and money
Another good question is, Does artificial intelligence need money? Let’s imagine a colony on Mars occupied by artificial intelligence alone. Would robots think that they still need to be paid? I think money is a very human thing. It has a very special role in the behavior of humans as well as strong relationships with reward, incentives and scarcity. These are all very human things.
What is needed for robots and AI is energy. We have to make the assumption that we’re talking about artificial intelligence here, and AI will work forever — as long as power is going to it. It doesn’t care for anything beyond that. It lacks an emotional component.
This may change when we get to artificial superintelligence or artificial general intelligence, where we may see and call AI a type of consciousness, but let’s stick with AI, which is what we know today. To put it simply, the notion of bartering and incentives doesn’t seem to play a role; I would say what is most important is an energy source.
If an AI has an amazing solar power grid on the moon with really efficient batteries, it’s going to work for hundreds of years — assuming it can be maintained without such issues.
However, there is a twist to this. Imagine for a moment that we’re already at the point of everyone riding in self-driving cars and every car is equal on the road. No vehicle gets priority over another in terms of lanes or intersections. Basically, the main priority is safety, so when self-driving vehicles come to an intersection where they’re going to collide, they have an interaction between each other to avoid the collision, but there’s no prioritization other than rules of the road.
Now imagine you get into one of these, and it’s an Uber of the future. The Uber has two fees: $10 for the trip and an optional $15. The difference is, if you pay that $15, your ride will get preference over lower-paying passengers. So let’s say you’ve paid an additional $15 and you get to this intersection — all other cars slow down. What happens is your car pays a premium to the other cars to slow down, giving yours the preferential route. So, these artificially intelligent cars — or, I’ve heard somebody call them “mobots” (mobile robots) — are going to start trading with each other, as they’re all going to have different pricing and a lot of different preferences. As they interact and have machine-to-machine communications, they will be sending cash or cryptocurrency.
It’s likely that, between the devices, preferential or less preferential treatment can be given or leased. Although it’s a little bit of a stretch in terms of the example above, robots paying for other robots to behave in certain ways based on different outcomes could become a reality. Ultimately, the question of risk in this example begins with: Where does the money go? Money trickles down to humans. Eventually, there will be shareholders or some organization incentivized to profit from this behavior.
Humans are basically the only creatures who create things symbolic of value, which then define our interactions.
Artificial intelligence and modern society
My next idea enters the paradigm of the post-truth society we live in. I worry a lot about the era we’re in right now. I like to think that in a few years, we’ll look back at this time and agree that it was a dark period.
Things started to get dark quickly in the years leading up to World War II in Europe. It didn’t even take decades. In the time that Adolf Hitler became known, moved through the ranks and became the chancellor, only three years passed. World War II raged for just six years, but in that brief window of time, Europe was destroyed almost beyond recognition and several millions of people were killed. Yet somehow out of that, we had the 1950s and the 1960s — periods of great optimism and change, liberalism and democracy, with a general positive hope for the future. At the time, anyone might have said that 1941 was the end or that there was no future for humanity, and yet here we are in the 21st century with a remarkably different world.
These periods of darkness in the last few hundred years are just periods of massive uncertainty and instability. Yet, things have to get sorted out. The world wars happened because of empires, and then the empires collapsed and were rejigged. Finally, countries started to set borders in the 1950s and 1960s, and things like the United Nations and European Union began to emerge.
Right now, the uncertainty is because suddenly, billions of people who never had a voice suddenly have one. Billions of people are connected and millions of them are moving to the middle class. Everybody wants a smartphone, a laptop, a car and a television, and more people want to meet. We are going through an incredible societal change while a dark cloud of climate change looms on the horizon.
If you look at the data, though — education, rights for women, rights for minorities, the rate of diseases, wars — everything is much better. We pay attention to the darkness, but we forget the bigger picture is actually quite positive.
I do think there are some very unstable, unpredictable things happening right now, a lot of which is generated by our massive use of networks and collaboration of technologies. This is one of the manifestations of our post-truth world, and I think my comments here might be more personal than generalized.
I think honesty and truth are essential not only to what it means to be human, but to creating stability in the world. So I am very concerned as both a citizen and as a human being that people who have power are not telling the truth, which encourages corporations and other entities to create untruths.
One specific example of this is deepfakes — videos powered by artificial intelligence, processing and visual technology showing people saying or doing things they never said or did. Soon, we won’t be able to differentiate between what is fake and what is reality. I think we’ll even see a movie in a few years that features actors that are no longer alive, perhaps with Marlon Brando and Marilyn Monroe — and they will seem completely and utterly real, we won’t not be able to tell with our eyes that it’s all fake.
Artificial intelligence is a powerful tool that can improve our lives greatly, and it is already doing this. At the same time, AI could have a nefarious impact on our society. How we will use this powerful technology is only up to us, humans.
This is part two of a multi-part series on digital future and technological innovations, read part one about quantum computing here.
This article is from an interview held by Kristina Lucrezia Cornèr with Dr. Jonathan Reichental. It has been condensed and edited.
The views, thoughts and opinions expressed here are the author’s alone and do not necessarily reflect or represent the views and opinions of Cointelegraph.
Dr. Jonathan Reichental is the CEO of Human Future, a global business and technology education, advisory and investment firm. He is the former chief information officer for the City of Palo Alto, and is a multiple-award-winning technology leader whose 30-year career has spanned both the public and private sectors.